Related problem: Entering URL into a field mapped to dc:identifier
When you are deciding which fields will contain URLs, please consider that one particular combination creates problems for many OAI
harvesters harvesting your collection (like MWDL).
If BOTH of these conditions exist for a field, it will create problems in how harvesters read the CONTENTdm-produced OAI
- The field is mapped to the Dublin Core Identifier field, AND
- The field starts with an Internet protocol designation, such as "http://" or "ftp://".
The problem this particular combination creates is that the harvester mistakenly reads this field as being the item URL, and it takes it instead of the reference URL that CONTENTdm creates, which is also mapped to dc:identifier. As a result, in MWDL, for example, when your item comes up in a search or browse, and the patron clicks on it, he/she is not taken to the item record. Instead, the browser attempts to go the URL in the field. If you entered a single URL in this field, the browser tries to open it; if you entered multiple URLs or other text after the first URL, the browser will be unable to resolve the link. Either way, the patron can't see your item page.
Any one of these will work:
- Change the mapping on this field to another Dublin Core field.
- Do not map the field at all, in which case it will not be harvested.
- Start the text in the field with anything other than an Internet protocol designation. Entering a small bit of text before the URL, like "Media location" or "FTP version", will work.
- 07 Jun 2008